Mobile health (mHealth) technologies are transforming healthcare delivery in low and middle-income countries by expanding access and improving continuity of care. In Ghana, however, adoption remains uneven, with most solutions constrained by poor connectivity, language barriers, and low digital literacy. This paper presents the design and evaluation of the National Health Access Platform (NHAP), an AI-enabled mHealth system developed to enhance healthcare access in Ghanaian underserved communities. NHAP integrates appointment scheduling, teleconsultation, multilingual health education, emergency response, and predictive analytics within a single platform. The system is built as an offline-first, low-bandwidth application, emphasizing inclusivity through local language support, intuitive interfaces, and secure data management. A cross-sectional evaluation involving 102 participants from all regions of Ghana showed strong usability outcomes, with mean scores of 4.5 for ease of navigation and 4.4 for likelihood to recommend. Overall, 87% of users found the app clear and easy to use, and 76% agreed it met their expectations. The findings demonstrate that combining artificial intelligence with human-centered design can produce scalable, context-aware digital health solutions that help bridge healthcare inequities in resource-limited settings.